The OSG Ctrl to right-click mapping was breaking this on Mac, but
changing that will break other things, so move the ruler feature from
‘ctrl left click’ to a simple ‘right click’ which ends up being the
same on Mac anyway.
Apply the device-pixel-ratio when calculating the initial position of
PUI dialogs, so they appear correctly (eg, centered) when running with
device-pixel-ratio != 1.0
The mapping from aircraft URI to name was wrong for local (non-package)
aircraft, as spotted by Thorsten Renk. Note other data (thumbnail) is
still incorrect, but another change I have pending will replace this
code anyway so only doing the simple fix for now.
Doing the option processing in Options::parseOption() has drawbacks:
- doesn't work well upon reset;
- doesn't work in the built-in launcher Additional Settings box.
Options::processOptions() is invoked both upon reset after the property
tree has been reset, and by the built-in launcher to process options
given in the Additional Settings box. This is not the case of
Options::parseOption() which is better for... parsing. :-)
Also use SGPath::fromLocal8Bit() to decode the path argument of --addon.
* Correct the filename case in the #include directive in PUICamera.cxx,
so that compilation succeeds on case-sensitive filesystems.
* Add the PU_SCROLL_UP_BUTTON and PU_SCROLL_DOWN_BUTTON #defines, as
they are missing from some systems which ship old versions of the PUI
library.
Move all PUI event and rendering into a custom camera, which can be
rendered via an FBO to account for display-resolution scaling (HiDPI).
Start wrapping PUI calls in #ifdefs so PUI can be disabled at compile
time; a run-time switch is trivial now but not implemented yet.
This makes it considerably easier to reference an item from the currently selected style without having to lookup the selected style index and then use that to select the appropriate element from the styles/ tree
Treat passing nil as clearing the speed / altitude restriction for
the leg. Also default the second arg to ‘at’, so calling these
methods with a single numerical arg is permitted.
Allows creating a new, inactive by empty Flightplan from Nasal. Can
also be used to load a flightplan:
var fp1 = createFlightplan();
var fp2 = createFlightplan("p3.xml");
Rewrite the position-init code for carrier starts, to precisely wait
on the carrier model being loaded, before proceeding with FDM init.
This allows the FDM to see the correct carrier model in the ground
cache, and hence avoids starting in the water.
To implement this, the CheckSceneryVisitor is used to force the carrier
model to be loaded while the splash-screen is visible.
In some testing, JSBsim sometimes passed extremely large values into
this function (larger than the radius of the earth). This causes
every tile on the planet to be scheduled if using the standard STG
terrain, effectively blocking startup.
This code allows us a chance to catch this case via logging, but does
not fix the underlying issues.
The problem is that the two aren't functionally equivalent; the Nasal version would fail silently - whereas the new C++ version gave a runtime error.
The immediate fix is therefore to simply remove the runtime error.
The bug fix is:
- if (!log.output) {
+ if ( !(*log.output) ) {
(i.e., testing the sg_ofstream instance instead of its address) and then
ensuring that the corresponding Log instance is removed from _logs and
destroyed.
The "destroy" part is made automatic by using std::unique_ptr instead of
raw pointers. This allows to simplify several areas of the code.
Don't provide custom definitions for the constructor and destructor of
FGLogger anymore, now that they don't need to do anything: IIRC, this
allows compilers to do some optimizations according to the C++ standard.
Since the paths of files overwritten by FGLogger come from the property
tree[1], they must be validated before we decide to write to these
files.
[1] Except for the "empty" case, which uses the default name
'fg_log.csv'. This file is deemed acceptable to overwrite in the
current directory, as the name is completely fixed and clearly
FG-specific.
Call fgInitAllowedPaths() right after Options::processOptions() (which,
among other things, determines $FG_ROOT and processes
--allow-nasal-read). This way, fgInitAllowedPaths() can be used in much
more code, such as when initializing subsystems.
Improve performance of Nasal properties access
by implementing the setValues/setChildren props.nas
functions in C++. Naive implementation effectively
copying the Nasal versions verbatim.
However, on a test suite writing 1000 Canvas paths,
results in a 50% reduction in runtime, and significant
reductions in memory occupancy.
On a much larger task (5500 paths) results in a 30%
reduction, probably due to Nasal GC.
Use base 26 numbering with letter-only digits for resource indices in
the C++ files generated by fgrcc. This is needed because, for instance,
'resource10' appears not to be a valid C++ variable name, mpfff...
Translations/en_US/FlightGear-nonQt.xlf is for a proper English
translation, where for instance "found %n airport(s)" would have two
plural forms, "found %n airport" and "found %n airports" (most
non-plural strings can be taken verbatim from the default translation,
and at this point there is no plural form at all yet).
As they are registered here, the files will have virtual paths such as:
/Translations/de/FlightGear-nonQt.xlf
/Translations/en_US/FlightGear-nonQt.xlf
etc.
for the EmbeddedResourceManager ('/' being the default virtual prefix).
This reduces the log noise in release builds, relating to AI ground-nets
with incomplete data, especially the commonly occurring ‘gate XYZ
doesn’t seem to have any routes associated with it’ message.
https://sourceforge.net/p/flightgear/codetickets/1974/
Updating the value takes immediately, as it did in 2017.1 and prior.
Additionally, passing —download-dir on the command-line is detected
and handled specially; the UI option is disabled, and no changes are
made to the value passed in. (The launcher does not override the value
with any value it has saved)
using --addon=/foo/bar does
* add /foo/bar/config.xml as propertyfile
* add /foo/bar to aircraft_paths to provide read-access
* sets property /addons/addon[n]/path = "/foo/bar"
* addons get initialized from addons.nas in FGDATA/Nasal
Change fgcommand to take an optional property tree root element.
This fixes the animation bindings to use the defined property tree root - to support multiplayer (or other) model that can bind to the correct part of the property tree.
Requires a corresponding fix in sg to allow the command methods to take an optional root parameter.
What this means is that when inside someone else's multiplayer model (e.g. backseat, or co-pilot), the multipalyer (AI) model will correctly modify properties inside the correct part of the property tree inside (/ai), rather than modifying the properties inside the same part of the tree as the non-ai model.
This means that a properly setup model will operate within it's own space in the property tree; and permit more generic multiplayer code to be written.
This is probably responsible for some of the pollution of the root property tree with MP aircraft properties.
Ground-bit is set manually (via the transponder GND knob position) or
automatically using a property referenced in instrumentation.xml.
Aircraft will only transmit these new properties when using the
multiplayer v2 protocol.
This adds console and message-box warnings, based upon aircraft
declaring the minimum FG version they support. A follow-up commit
will extend the launcher UI to warn the user about this in a nicer
way.
This is a headless mode, designed to be invoked from an installer, not
used directly by users. It doesn’t touch the ‘normal’ installation, but
rather removes the other files FG typically creates or downloads.
When building with MSVC, use the CMAKE_MSVCIDE_RUN_PATH variable to
prepend ${MSVC_3RDPARTY_ROOT}/${MSVC_3RDPARTY_DIR}/bin to the PATH in
order to (hopefully) allow fgrcc to find the libraries it needs. We may
need to add something similar for SimGear---will see.
The use of CMAKE_MSVCIDE_RUN_PATH and other ways to address this problem
are discussed at
<https://stackoverflow.com/questions/28533012/how-to-set-runtime-path-for-cmake-custom-command-on-windows>.
${CMAKE_SOURCE_DIR}/src/EmbeddedResources/FlightGear-resources.xml
(currently empty) is automatically "compiled" into
${CMAKE_BINARY_DIR}/src/EmbeddedResources/FlightGear-resources.[ch]xx by
fgrcc inside the build directory. These files are incorporated into the
FlightGear build (FlightGear-resources.cxx is linked into FlightGear).
When the XML embedded resource declaration file added here,
FlightGear-resources.xml, is compiled, fgrcc is passed the
--root=${CMAKE_SOURCE_DIR} option, so that files referred to in
FlightGear-resources.xml are looked up relatively to the root directory
of the FlightGear repository. One could use a second XML embedded
resource declaration file compiled with a different --root option to
grab files from FGData, for instance. I would name such a file
FGData-resources.xml to be consistent with the current naming scheme.
Note: this --root option applies to the paths of real files. Don't
confuse it with the 'prefix' attribute of <qresource> elements
inside XML resource declaration files (such as
FlightGear-resources.xml), which applies to the virtual path of
each resource defined beneath.
The commands in src/Main/CMakeLists.txt ensure that
FlightGear-resources.xml is recompiled with fgrcc whenever it is
changed, and obviously also when FlightGear-resources.cxx or
FlightGear-resources.hxx is missing. However, CMake doesn't know how to
parse fgrcc XML resource declaration files, therefore when a resource is
modified but the XML file it is declared in is not (here,
FlightGear-resources.xml), you have to trigger yourself a recompilation
of the XML resource declaration file to see the new resource contents
inside FlightGear. The easiest ways to do so are:
- either update the timestamp of the XML resource declaration file;
- or remove one or both of the generated files
(FlightGear-resources.cxx and FlightGear-resources.hxx here).
The EmbeddedResourceManager is created in fgMainInit() just after
Options::processOptions() set the language that was either requested by
the user or obtained from the system (locales). Resources from
FlightGear-resources.cxx are added to it, after which
EmbeddedResourceManager::selectLocale() is called with the user's
preferred locale (obtained with FGLocale::getPreferredLanguage()).
Upon reset (fgStartNewReset()), EmbeddedResourceManager::selectLocale()
is called in a similar way after Options::processOptions(), however in
this case the EmbeddedResourceManager instance doesn't have to be
recreated.
- remove use of boost in src/Main/locale.cxx;
- add missing header <cstring> for std::strlen();
- replace NULL with nullptr;
- fix some broken indentation;
- other small readability improvements.
This function returns the preferred "locale"[1] according to user choice
and/or settings (i.e., it is influenced by --language if passed,
otherwise by current locale/system settings). The return value never has
an encoding part. It is the empty string if nothing could be found,
otherwise should look like fr_BE or it_IT.
[1] "language" term used in the function name for consistency with the
existing and related FGLocale::selectLanguage().
Windows and Mac implementations return a string without any encoding
specifier -> remove this specifier directly in the Unix/Linux
implementation for consistency.
Also do some small refactoring with the new static method
FGLocale::removeEncodingPart(). Slight difference with the previous
algorithm: if a '.' is found in the given locale spec, we assert() that
it is not the first character. The previous code in
FGLocale::findLocaleNode() used to consider such weird locale specs
starting with a dot as normal locale specs without any encoding part.
Note: the same change could be done where FGLocale::findLocaleNode()
looks for an underscore in order to prepare for the fallback
search (e.g., 'fr' after not finding translations for 'fr_FR').
The wake induced force and moment are computed as an external reaction to the JSBSim model. For these force and moment to be accounted for by JSBSim, the following external reaction needs to be added to FDM XML definition:
<external reaction>
<force name="ai-wake">
...
</force>
<moment name="ai-wake">
...
</moment>
</external reaction>
This is similar to how the hook and wire feature is modeled in JSBSim.
Wake computations are now performed for all AI aircrafts within a range lower or equal to the value indicated by the property /fdm/ai-wake/max-radius-nm.
These computations are triggered by the property /fdm/ai-wake/enabled (it is disabled by default).
The result of the wake computations is not yet used by the FDMs so do not expect the user aircraft to react to the AI wake.
AI wake code is still dead code except that it is now compiled with FG.
Input data for wake computations are extracted from the performance database. The data must be specified as follows (values are for illustration only) :
<geometry>
<wing>
<span-ft> 100. </span-ft>
<chord-ft> 12. <chord-ft>
</wing>
<weight-lbs> 90000. </weight-lbs>
</geometry>
At the moment, this is dead code: only the tests are compiled. FG is still compiled without this code.
A new directory is created that contains all the numerical computations made to estimate the wake induced by AI aircrafts. This is based on the venerable Vortex Lattice Method (VLM) which was all the rage in the 60's Computational Fluid Dynamics (CFD).
Even though quite old, the method is relevant to compute aircrafts wake in real time since 3D Navier-Stokes (NS3D) is out of reach for real time computations even with modern multicore personal computers and their GPUs.
Use owning pointers in GraphicsWinQt to own the GLwindow and QOGLContext,
so shutdown is more orderly, especially when driven from the OSG or
window-manager, as opposed to an ‘internal’ quit.
In order to compute the AI aircrafts wake, the FDM will need access to the list of AI aircrafts as well as being able to compute their range to discard aircrafts which are too far.
In the thread at
<https://sourceforge.net/p/flightgear/mailman/flightgear-devel/thread/87tw7sm5uw.fsf%40frougon.crabdance.com/#msg35673402>,
it was decided that alignLocaliserWithRunway() was a bad idea overall,
and should just be removed. The feature was already disabled by default
in recent releases via defaults.xml (since FlightGear 2017.1, precisely
FGData commit ea76305e42fd245e79a2b5bb569c8c61161f5721). Remove the code
entirely now.
The properties /sim/navdb/localizers/auto-align and
/sim/navdb/localizers/auto-align-threshold-deg have no effect anymore,
you may remove them from your configs if you somehow set them.
This goes with FGData commit 6dd9f6e8d962b00a1ccddc6af487c404cc935b92.
In case someone wanted to reuse the removed code, please consider
Szymon's proposed change at
<https://sourceforge.net/p/flightgear/flightgear/merge-requests/76/>,
which was:
,----
| Fixed bug in alignLocaliserWithRunway when building NavCacheDb
|
| The position of the localizer was converted to a distance from
| the displaced threshold, and then back to the posistion on the
| centerline, but the origin of the second computation was
| non-displaced threshold.
|
| --- a/src/Navaids/navdb.cxx
| +++ b/src/Navaids/navdb.cxx
| @@ -104,8 +104,8 @@
| void alignLocaliserWithRunway(FGRunway* rwy, const string& ident, SGGeod& pos, double& heading)
| {
| assert(rwy);
| - // find the distance from the threshold to the localizer
| - double dist = SGGeodesy::distanceM(pos, rwy->threshold());
| + // find the distance from the (non-displaced) threshold to the localizer
| + double dist = SGGeodesy::distanceM(pos, rwy->geod());
|
| // back project that distance along the runway center line
| SGGeod newPos = rwy->pointOnCenterline(dist);
`----
There was a couple of typos in the routine. This code will be used by the AI wake feature.
Also added a test for the matrix code to make sure they will be no regression.
External moments can now be specified in addition to external forces. This feature is needed for the AI wakes.
Headers clean up to reduce the amount of recompilation when modifying an header file.
Print an error message and exit if --{enable,disable}-enhanced-lighting
or --adf are used (those deprecated options will be removed in a future
version of FlightGear).
As spotted by Thorsten, reduce user annoyance by asking about FGData
when the built-in data is acceptable. (Advanced users can use the
launcher or other options to change FGdata as normal).
* Skip the gears when located over water. They will be skipped by the ground reactions anyway making the trim irrelevant.
* Ignore the ground bumpiness during the trim. The trim algorithm seems not to
be robust enough to handle that. This does not make much difference to the
converged solution anyway since the bumpiness is generally low.
* Fixed the WOW status for contacts over water (was 'false' is now 'true').
* All the gear/contact data are now properly reset when WOW==false. The reset
code is now common to all the case where WOW is false.
When the user specified an MP server without providing a port, the
launcher would pass a malformed command line, leading the the simulator
bailing out.
(Should go to the release branch)
Add defensive logic to ensure that a null pointer isn't dereferenced. This happened to me once or twice during a reinit (Location/Position In Air to the current position). It may be related to the model, but in any case the code needs to ensure that the static cast failure is handled correctly.
2017.2 in compatible mode was not compatible with 2017.1.x because the V1 packet had V2 properties in it; the protocol version and the new launchbar encoding - both of which should not be transmitted. This only affected 2017.1 because it discards the rest of the packet once an unknown property is encountered (to increase reliability).
Add support for per property encoding.
- encode_for_transmit - method that will convert from and to the packet for the value.
- decode_received - decodes received data
The options.cxx code is not ready to handle recursive use of --config
(for config files). Instead of failing in an ugly way, abort with a
clear error message in such situations. See discussion at
<https://sourceforge.net/p/flightgear/mailman/message/35838852/>.
Note: it *is* possible to load XML PropertyList files from config files,
so --config is not entirely "banned" from config files.
+ add missing include
New Nasal method get_cart_ground_intersection
Returns where the given position in the specified direction will intersect with the ground. Returns whether or not a certain position and direction pair intersect with the ground, and if so the intersection point.
Useful for radars, terrain avoidance (GPWS), etc.
Input parameters:
1. vec3d(x,y,z) position
2. vec3d(x,y,z) direction
Returns nil or geod hash (lat:rad,lon:rad,elevation:Meters) intersection
Example Usage:
var end = geo.Coord.new(start);
end.apply_course_distance(heading, speed_horz_fps*FT2M);
end.set_alt(end.alt() - speed_down_fps*FT2M);
var dir_x = end.x() - start.x();
var dir_y = end.y() - start.y();
var dir_z = end.z() - start.z();
var xyz = { "x":start.x(), "y" : start.y(), "z" : start.z() };
var dir = { "x":dir_x, "y" : dir_y, "z" : dir_z };
var geod = get_cart_ground_intersection(xyz, dir);
if (geod != nil) {
end.set_latlon(geod.lat, geod.lon, geod.elevation);
var dist = start.direct_distance_to(end)*M2FT;
var time = dist / speed_fps;
setprop("/sim/model/radar/time-until-impact", time);
}
-----------
Nasal method aircraftToCart : This allows easily computing offsets in aircraft-relative coordinates, and converting to global cartesian (ECEF) reference frame.
This has the advantage, according to my testing on Linux, that core
files obtained after a crash now point to the crashing thread again,
when one starts 'gdb' on the core file and runs the 'bt' command.
Apparently, when using kill(), the signal is seen as coming from the
outside and gdb's 'bt' command points to the wrong thread in general
when debugging using a core file (when debugging "live", gdb intercepts
the signal even before FG's signal handler is started).
See discussion starting at
<https://sourceforge.net/p/flightgear/mailman/message/35833221/>.
Add an additional visibility flag to the menubar implementations,
conditional on whether or not the menubar overlaps the window content.
(I.e for PUI but not Cocoa). This flag is linked to a new property
/sim/menubar/overlap-hide, which the renderer drives off the splash-
screen visibility.
On Windows, use:
- QtFileDialog if FG was built with Qt support;
- PUIFileDialog otherwise.
Behavior on other platforms is unchanged. This change is motivated by
the fact that some Windows users have reported[1][2] weird,
non-deterministic behavior of WindowsFileDialog and unfortunately, no
one seems to be willing and able to fix the problem. The Qt
implementation comes for free and should be quite robust. Of course, if
someone wants to maintain the WindowsFileDialog class again, the change
can be reverted.
See discussion at [3].
[1] https://forum.flightgear.org/viewtopic.php?f=25&t=31945
[2] https://sourceforge.net/p/flightgear/mailman/message/35761650/
[3] https://sourceforge.net/p/flightgear/mailman/message/35759819/
This fixes handling of non-ASCII splashscreen text for me (Debian
GNU/Linux). The XML input files don't technically *have* to be encoded
in UTF-8, as long as they properly declare the encoding and it is
supported by Expat. Of course, we prefer UTF-8 nowadays.
With this commit, startup tips and splash screen progress strings
("Loading scenery", "Initializing subsystems", etc.) can at last be
written correctly in languages that need non-ASCII characters.
The Expat doc is unfortunately unclear on its *output* encoding, saying
the following (expat.h):
The characters are passed exactly as they were in the XML document
except that they will be encoded in UTF-8 or UTF-16.
The only relevant header I can see in SimGear is
3rdparty/expat/sg_expat_external.h, which has interesting stuff around
XML_UNICODE, however it doesn't seem to take position.
I fear that whether UTF-8 or UTF-16 is used for Expat's output (and thus
for what easyxml.cxx gives us) depends on how it was compiled. Let's
hope everyone has it compiled for UTF-8 output...
In commit 15525aab58, a layer with
lapse=0.0 was added to ISA_def (atmosphere model). However, the last
layer *must* have lapse == -1.0, otherwise the code in PT_vs_hpt()
doesn't know when it's on the last element of ISA_def, and does an
illegal memory access with (pp+1)->height. This is why -1.0 is the
default value for lapse ('l') in the ISA_layer constructor.
Fix: add the -1.0 terminator to the last element of ISA_def.
When the test:
if (i == activeTraffic.end() || (activeTraffic.empty()))
was true, the iterator named 'current' was uninitialized and later
dereferenced.
Fix:
- when the previously-mentioned test is true, return;
- initialize 'current' only when it is really needed (i.e., later and
after the test), and since we don't need it for iterating, make it
an FGTrafficRecord&.
- Don't ignore pushbackRoute="0".
- Stricter parsing with precise log messages when the input is
incorrect.
- Add missing includes in src/Airports/dynamicloader.cxx.
See <https://sourceforge.net/p/flightgear/mailman/message/35788373/> for
the discussion about this change.
Before SimGear commit a962c90b30f36575d01162b64471fa77473237a0,
SGPath::pathListSep was a char in static memory that was not necessarily
followed by '\0'. As a consequence, using &SGPath::pathListSep as a
C-style string could result in a string containing the correct separator
*plus* whatever followed in memory until the first null byte...
SimGear commit a962c90b30 changes this situation by making
SGPath::pathListSep an array of two const chars: the path list separator
followed by a '\0'.
This commit simply adapts FlightGear to this change, which fixes a
couple of bugs where the separator was used, mainly unneeded NavCache
rebuilds due to the "apt.dat", "fix.dat" and "nav.dat" properties in the
SQLite database containing the correct paths separated by a possibly
incorrect separator string (there was no alteration of the cache
contents as far as I can tell, since the db property is only used to
check if the lists of apt.dat, fix.dat and nav.dat files have changed).
Some buggy *.groundnet.xml files (as KSEA currently on TS) define the
pushback hold point for some parking positions as a node on a runway.
In this case, this the pushback hold point for parking
'North_Cargo_Ramp', defined as node 5344 in
Airports/K/S/E/KSEA.groundnet.xml, which is defined twice (second error),
first as:
<node index="5344" lat="N47 27.774559" lon="W122 18.465257" isOnRunway="1" holdPointType="PushBack" />
and then as:
<node index="5344" lat="N47 27.725747" lon="W122 18.159649" isOnRunway="1" holdPointType="PushBack" />
(due to code in flightgear/src/Airports/dynamicloader.cxx, it should be
the second one that wins, which is not on a runway but on apron in the
north cargo area)
As a consequence, when this gate is selected for an AI aircraft, the
pushback route has only one node (since the pushback hold point is then
the closest point to itself supposedly on runway!), and the
corresponding FGTaxiRoute instance has an empty 'routes' member
variable, which FGTaxiRoute::next() doesn't handle gracefully
(segfault).
It may be that an additional check/change could be desirable in
FGTaxiRoute::next() in such a case (one node and obviously no route in
the FGTaxiRoute instance), however I'm not sure how Durk wants this case
to be handled, since FGTaxiRoute::next() seems to iterate on nodes.
This fixes the bug reported at:
https://forum.flightgear.org/viewtopic.php?p=308397#p308397 and
https://sourceforge.net/p/flightgear/mailman/message/35776552/
Thanks to yanfiz and wkitty42 for the report, and to gooneybird for
inspecting the groundnet file.
This is because FGTaxiNode::ident() is generally (always?) an empty
string for FGTaxiNode instances. This concerns the:
unreferenced groundnet node: ...
warning. Also remove one tiny use of boost.
We now show paths in ‘view command line’ and set them through the
standard mechanism. Re-ordering the paths also notifies the rest of
the system correctly.
Code and tests to demonstrate migrating of older auto-save files, with
blacklisting support to exclude properties. Disabled pending agreement
on the required blacklisting values.
Some pieces of code such as fgMainInit() and, by cascading effect,
fgInitHome(), were careful to return a meaningful value indicating
success or error, however the main() function in src/Main/bootstrap.cxx
ignored it royally so far.
main() now returns:
- EXIT_FAILURE if fgMainInit() or fgviewerMain() throws an exception;
- whatever said function returns otherwise.
- Rename fatalMessageBox() to fatalMessageBoxWithoutExit(). This should
prevent the kind of bug that prompted this set of changes: someone
calling fatalMessageBox(), assuming the program would stop at that
point, whereas in reality it did not.
- Add new function fatalMessageBoxThenExit(). This is not vital of
course, but allows one to spare one line here and there and to apply
the DRY principle for such fatal exits.
- Replace every existing call to fatalMessageBox() with one or the other
of the two new functions. Improve formatting along the way. This
fixes a few bugs of the kind explained above.
This reverts commit 9e6a3ebc6b ("Make
fatalMessageBox() end with std::abort() and declare it [[noreturn]]").
After reflexion, it seems better to let fatalMessageBox() return,
because there is existing code that appears to be relying on this aspect
to do some work after having called fatalMessageBox() (cf. main() in
bootstrap.cxx). Also, the way of exiting from fatalMessageBox() after
commit 9e6a3ebc6b (std::abort()) was probably too brutal for a
controlled exit---as opposed to a terminate handler.
Basically, this is because fatalMessageBox() is only safe to call from
the GUI thread, however it seems fg_terminate() can be called from any
thread (according to C++11 semantics). Additionally, fatalMessageBox()
typically requires some work to happen in the GUI thread (event loop) in
order to display something, but we can't realistically expect this while
running a terminate handler just before the program dies.
See messages around
<https://sourceforge.net/p/flightgear/mailman/message/35775803/> for a
discussion of this subject.
+ Minor header cleanup (<locale.h> replaced with <clocale>, etc.)
/sim/multiplay/protocol-version is either 1 or 2 and controls how packets are sent. V2 packets will only have the (motioninfo) basic properties visible to older clients.
New string encoding that is efficient.
Support short int encoding (pack a property and value into 4 bytes).
Allow properties to be transmitted using a different encoding to the property type in the tree.
Support scaled floats; most of the floats we transmit are small and thus can fit into a scaled short.
V2 protocol uses transmit so most properties are either scaled floats or short ints.
Allow the client to request a larger visibility range by setting /sim/multiplay/visibility-range-nm. This will transmit in the repurposed header field ReplyAddress - which has been renamed to RequestedRangeNm. This will require support from fgms to actually do anything.
Extra debugging options. The most useful (for aircraft developers) is the loopback bit, as this allows model multiplay testing without running two instances.
Update property /sim/multiplay/last-xmit-packet-len with the size of the packet transmitted
Debug level bits in property /sim/multiplay/debug-level
bit 1 - loopback (show your own model as an MP model)
bit 2 - dump outgoing packets
bit 3 - dump incoming packets
bit 4 - hexdump outgoing packets
/sim/multiplay/protocol-version is either 1 or 2 and controls how packets are sent. V2 packets will only have the (motioninfo) basic properties visible to older clients.
New string encoding that is efficient.
Support short int encoding (pack a property and value into 4 bytes).
Allow properties to be transmitted using a different encoding to the property type in the tree.
Support scaled floats; most of the floats we transmit are small and thus can fit into a scaled short.
V2 protocol uses transmit so most properties are either scaled floats or short ints.
Allow the client to request a larger visibility range by setting /sim/multiplay/visibility-range-nm. This will transmit in the repurposed header field ReplyAddress - which has been renamed to RequestedRangeNm. This will require support from fgms to actually do anything.
Extra debugging options. The most useful (for aircraft developers) is the loopback bit, as this allows model multiplay testing without running two instances.
Update property /sim/multiplay/last-xmit-packet-len with the size of the packet transmitted
Debug level bits in property /sim/multiplay/debug-level
bit 1 - loopback (show your own model as an MP model)
bit 2 - dump outgoing packets
bit 3 - dump incoming packets
bit 4 - hexdump outgoing packets
We probably need a warning for cross-aircraft paths, but leaving that
for a separate change since I’m worried it will warn on MP aircraft.
Maybe better checked in the Python scripts than in the app?
Compile a useful subset of FG as a shared library, and add two basic
uses of this to exercise some Flightplan / RoutePath / navaid
functions.
The test framework can/will be expanded incrementally from here, this
is just a starting point.
